Predicting bacteraemia in validated models--a systematic review

Added Value of Shaking Chills for Predicting Bacteremia in Patients with Suspected Infection

BACKGROUND

Detailed grading of chills is more useful for diagnosing bacteremia than simply classifying the presence or absence of chills. However, its value added to other clinical information has not been evaluated.

OBJECTIVE

To evaluate the value of adding chills grading to other clinical information compared to simply noting the presence or absence of chills for predicting bacteremia in patients with suspected infection.

DESIGN

Prospective observational study.

PARTICIPANTS

Adult patients admitted to two acute-care hospitals with suspected infection from April 2018 to March 2019.

MAIN MEASURES

Two types of categorization for chills were applied: "presence" or "absence" (dichotomized chills); and "no chills", "mild/moderate chills", and "shaking chills" (trichotomized chills). Three multivariable logistic regression models incorporating each of dichotomized chills, trichotomized chills, and C-reactive protein (CRP) with other clinical information were developed and compared. To assess the potential consequences of using each model to identify patients with high risk of bacteremia (i.e., requiring prompt intervention), we applied a cut-off point of an estimated probability of 60%. The number of patients with bacteremia correctly identified by each model was compared.

KEY RESULTS

Among the 2,013 patients, 327 (16.2%) were diagnosed with bacteremia. The three models showed comparable discrimination and calibration performance. At the 60% cut-off, the dichotomized chills model correctly identified 11 patients (3.4% [95% confidence interval (CI) 1.9-3.4] of patients with bacteremia). The trichotomized chills model and CRP model correctly identified an additional 15 patients (4.6% [95% CI 2.8-7.4]) and 2 patients (0.6% [95% CI 0.1-2.3]) with bacteremia, respectively.

CONCLUSIONS

Differentiating shaking chills in comparison with dichotomized chills for predicting bacteremia allowed the correct identification of an additional 4.6% of patients with bacteremia. Detailed grading of chills can be assessed without additional time, cost, or burden on patients and can be recommended in the routine history taking.

Assessing sepsis-induced immunosuppression to predict positive blood cultures

Introduction

Bacteremia is a life-threatening condition that can progress to sepsis and septic shock, leading to significant mortality in the emergency department (ED). The standard diagnostic method, blood culture, is time-consuming and prone to false positives and false negatives. Although not widely accepted, several clinical and artificial intelligence-based algorithms have been recently developed to predict bacteremia. However, these strategies require further identification of new variables to improve their diagnostic accuracy. This study proposes a novel strategy to predict positive blood cultures by assessing sepsis-induced immunosuppression status through endotoxin tolerance assessment.

Methods

Optimal assay conditions have been explored and tested in sepsis-suspected patients meeting the Sepsis-3 criteria. Blood samples were collected at ED admission, and endotoxin (lipopolysaccharide, LPS) challenge was performed to evaluate the innate immune response through cytokine profiling.

Results

Clinical variables, immune cell population biomarkers, and cytokine levels (tumor necrosis factor [TNFα], IL-1β, IL-6, IL-8, and IL-10) were measured. Patients with positive blood cultures exhibited significantly lower TNFα production after LPS challenge than did those with negative blood cultures. The study also included a validation cohort to confirm that the response was consistent.

Discussion

The results of this study highlight the innate immune system immunosuppression state as a critical parameter for sepsis diagnosis. Notably, the present study identified a reduction in monocyte populations and specific cytokine profiles as potential predictive markers. This study showed that the LPS challenge can be used to effectively distinguish between patients with bloodstream infection leading to sepsis and those whose blood cultures are negative, providinga rapid and reliable diagnostic tool to predict positive blood cultures. The potential applicability of these findings could enhance clinical practice in terms of the accuracy and promptness of sepsis diagnosis in the ED, improving patient outcomes through timely and appropriate treatment.

Leveraging explainable artificial intelligence for early prediction of bloodstream infections using historical electronic health records

Bloodstream infections (BSIs) are a severe public health threat due to their rapid progression into critical conditions like sepsis. This study presents a novel eXplainable Artificial Intelligence (XAI) framework to predict BSIs using historical electronic health records (EHRs). Leveraging a dataset from St. Olavs Hospital in Trondheim, Norway, encompassing 35,591 patients, the framework integrates demographic, laboratory, and comprehensive medical history data to classify patients into high-risk and low-risk BSI groups. By avoiding reliance on real-time clinical data, our model allows for enhanced scalability across various healthcare settings, including resource-limited environments. The XAI framework significantly outperformed traditional models, particularly with tree-based algorithms, demonstrating superior specificity and sensitivity in BSI prediction. This approach promises to optimize resource allocation and potentially reduce healthcare costs while providing interpretability for clinical decision-making, making it a valuable tool in hospital systems for early intervention and improved patient outcomes.

Procalcitonin for safe reduction of unnecessary blood cultures in the emergency department: Development and validation of a prediction model

OBJECTIVES

Blood cultures (BCs) are commonly ordered in emergency departments (EDs), while a minority yields a relevant pathogen. Diagnostic stewardship is needed to safely reduce unnecessary BCs. We aimed to develop and validate a bacteremia prediction model for ED patients, with specific focus on the benefit of incorporating procalcitonin.

METHODS

We included adult patients with suspected bacteremia from a Dutch ED for a one-year period. We defined 23 candidate predictors for a "full model", of which nine were used for an automatable "basic model". Variations of both models with C-reactive protein and procalcitonin were constructed using LASSO regression, with bootstrapping for internal validation. External validation was done in an independent cohort of patients with confirmed infection from 71 Spanish EDs. We assessed discriminative performance using the C-statistic and calibration with calibration curves. Clinical usefulness was evaluated by sensitivity, specificity, saved BCs, and Net Benefit.

RESULTS

Among 2111 patients in the derivation cohort (mean age 63 years, 46% male), 273 (13%) had bacteremia, versus 896 (20%) in the external cohort (n = 4436). Adding procalcitonin substantially improved performance for all models. The basic model with procalcitonin showed most promise, with a C-statistic of 0.87 (0.86-0.88) upon external validation. At a 5% risk threshold, it showed a sensitivity of 99% and could have saved 29% of BCs while only missing 10 out of 896 (1.1%) bacteremia patients.

CONCLUSIONS

Procalcitonin-based bacteremia prediction models can safely reduce unnecessary BCs at the ED. Further validation is needed across a broader range of healthcare settings.

A machine learning model for the early diagnosis of bloodstream infection in patients admitted to the pediatric intensive care unit

Bloodstream infection (BSI) is associated with increased morbidity and mortality in the pediatric intensive care unit (PICU) and high healthcare costs. Early detection and appropriate treatment of BSI may improve patient's outcome. Data on machine-learning models to predict BSI in pediatric patients are limited and neither study included time series data. We aimed to develop a machine learning model to predict an early diagnosis of BSI in patients admitted to the PICU. This was a retrospective cohort study of patients who had at least one positive blood culture result during stay at a PICU of a tertiary-care university hospital, from January 1st to December 31st 2019. Patients with positive blood culture results with growth of contaminants and those with incomplete data were excluded. Models were developed using demographic, clinical and laboratory data collected from the electronic medical record. Laboratory data (complete blood cell counts with differential and C-reactive protein) and vital signs (heart rate, respiratory rate, blood pressure, temperature, oxygen saturation) were obtained 72 hours before and on the day of blood culture collection. A total of 8816 data from 76 patients were processed by the models. The machine committee was the best-performing model, showing accuracy of 99.33%, precision of 98.89%, sensitivity of 100% and specificity of 98.46%. Hence, we developed a model using demographic, clinical and laboratory data collected on a routine basis that was able to detect BSI with excellent accuracy and precision, and high sensitivity and specificity. The inclusion of vital signs and laboratory data variation over time allowed the model to identify temporal changes that could be suggestive of the diagnosis of BSI. Our model might help the medical team in clinical-decision making by creating an alert in the electronic medical record, which may allow early antimicrobial initiation and better outcomes.

The Yield of One vs. Two Blood Cultures in Children: Under-Detection and Over-Testing

We aimed to determine whether obtaining two blood cultures (BCs) instead of one improved the detection of bloodstream infections (BSIs) in children. For this descriptive study, we used surveillance data collected in 2019-2021 from all Israeli hospitals serving children. The sample included 178,702 culturing episodes. One BC was taken in 90.1% of all episodes and 98.2% of episodes in the emergency department. A true pathogen was detected in 1687/160,964 (1.0%) of single-culture episodes and 1567/17,738 (8.9%) of two-culture episodes (p < 0.001). The yield was significantly different even when considering only the first BC in two-culture episodes: 1.0% vs. 7.5%. Among 1576 two-culture episodes that were positive for a true pathogen, the pathogen was detected only in the second culture in 252 (16.0%). We estimated that if a second culture had been taken in all episodes, an additional 343 BSIs by a true pathogen would have been detected. Among 1086 two-culture episodes with commensal bacteria, the second BC was sterile in 530 (48.8%), suggesting contamination. A commensal was isolated in 3094/4781 (64.7%) positive single-culture episodes, which could represent BSI or contamination. The yield of a single BC bottle was low, reflecting both lower sensitivity of a single bottle and the taking of single bottles in patients with a low probability of BSI.

Derivation and External Validation of the Ottawa Bloodstream Infection Model for Acutely Ill Adults

BACKGROUND

Knowing the probability that patients have a bloodstream infection (BSI) could influence the ordering of blood cultures and interpretation of their preliminary results. Many previous BSI probability models have limited applicability and accuracy. This study used currently recommended modeling techniques and a large sample to derive and validate the Ottawa BSI Model.

METHODS

At a tertiary care teaching hospital, we retrieved a random sample of 4180 adults having blood cultures in our emergency department or during the initial 48 h of the encounter. Variable selection was based on clinical experience and a systematic review of previous model performance. Model performance was measured in a temporal external validation group of 4680 patients.

RESULTS

A total of 327 derivation patients had a BSI (8.0%). BSI risk increased with increased number of culture sets (2 sets: adjusted odds ratio [aOR] 1.52 [1.10-2.11]; 3 sets: 1.99 [0.86-4.58]); with indwelling catheter (aOR 2.07 [1.34-3.20); with increasing temperature, heart rate, and neutrophil-lymphocyte ratio; and with decreasing systolic blood pressure, platelet count, urea-creatinine ratio, and estimated glomerular filtration rate. In the temporal external validation group, model discrimination was good (c-statistic 0.71 [0.69-0.74]) and calibration was very good (integrated calibration index .016 [.010-.024]). Exclusion of validation patients with acute SARS-CoV-2 infection improved discrimination slightly (c-statistic 0.73 [0.69-0.76]).

CONCLUSIONS

The Ottawa BSI Model uses commonly available data to return an expected BSI probability for acutely ill patients. However, it cannot exclude BSI and its complexity requires computational assistance to use.

External validation of ID-BactER and Shapiro scores for predicting bacteraemia in the emergency department

Introduction

The blood culture positivity rate in the emergency department (ED) is <20%; however, the mortality associated with Community-acquired bacteraemia (CAB) is as high as 37.8%. For this reason, several models have been developed to predict blood culture positivity for the diagnosis of CAB.

Objective

To validate two bacteraemia prediction models in a high-complexity hospital in Colombia.

Design

External validation study of the ID-BactER and Shapiro scores based on a consecutive cohort of patients who underwent blood culture within 48 h of ED admission.

Methods

Scale calibration was assessed by comparing expected and observed events (calibration belt). Discriminatory ability was assessed by area under the ROC curve (AUC-ROC).

Results

We included 1347 patients, of whom 18.85% were diagnosed with CAB. The most common focus of infection was the respiratory tract (36.23%), and the most common microorganism was Escherichia coli (52.15%). The Shapiro score underestimated the risk in all categories and its discriminatory ability was poor (AUC 0.68 CI 95% 0.64-0.73). In contrast, the ID-BactER score showed an adequate observed/expected event ratio of 1.07 (CI 0.85-1.36; p = 0.018) and adequate calibration when expected events were greater than 20%, in addition to good discriminatory ability (AUC 0.74 95% CI 0.70-0.78).

Conclusion

The Shapiro score is not calibrated, and its discriminatory ability is poor. ID-BactER has an adequate calibration when the expected events are higher than 20%. Limiting blood culture collection to patients with an ID-BactER score ⩾4 could reduce unnecessary blood culture collection and thus health care costs.

A clinical decision rule to rule out bloodstream infection in the emergency department: retrospective multicentric observational cohort study

BACKGROUND

We aimed to identify patients at low risk of bloodstream infection (BSI) in the ED.

METHODS

We derived and validated a prediction model to rule out BSI in the ED without the need for laboratory testing by determining variables associated with a positive blood culture (BC) and assigned points according to regression coefficients. This retrospective study included adult patients suspected of having BSI (defined by at least one BC collection) from two European ED between 1 January 2017 and 31 December 2019. The primary end point was the BSI rate in the validation cohort for patients with a negative Bacteremia Rule Out Criteria (BAROC) score. The effect of adding laboratory variables to the model was evaluated as a second step in a two-step diagnostic strategy.

RESULTS

We analysed 2580 patients with a mean age of 64 years±21, of whom 46.1% were women. The derived BAROC score comprises 12 categorical clinical variables. In the validation cohort, it safely ruled out BSI without BCs in 9% (58/648) of patients with a sensitivity of 100% (95% CI 95% to 100%), a specificity of 10% (95% CI 8% to 13%) and a negative predictive value of 100% (95% CI 94% to 100%). Adding laboratory variables (creatinine ≥177 µmol/L (2.0 mg/dL), platelet count ≤150 000/mm3 and neutrophil count ≥12 000/mm3) to the model, ruled out BSI in 10.2% (58/570) of remaining patients who had been positive on the BAROC score. The BAROC score with laboratory results had a sensitivity of 100% (95% CI 94% to 100%), specificity of 11% (95% CI 9% to 14%) and negative predictive value of 100% (95% CI 94 to 100%). In the validation cohort, there was no evidence of a difference in discrimination between the area under the receiver operating characteristic for BAROC score with versus without laboratory testing (p=0.6).

CONCLUSION

The BAROC score safely identified patients at low risk of BSI and may reduce BC collection in the ED without the need for laboratory testing.

Procalcitonin and Risk Prediction for Diagnosing Bacteremia in Hospitalized Patients: A Retrospective, National Observational Study

Bacteremia is associated with significant morbidity and mortality. Timely, appropriate therapy may improve clinical outcomes, and therefore, determining which patients benefit from more comprehensive diagnostic strategies (i.e., direct specimen testing) could be of value. We performed an assessment of procalcitonin (PCT) and clinical characteristics in the discrimination of bacteremic hospitalizations. We analyzed 71,105 encounters and 14,846 visits of patients with bacteremia alongside 56,259 without an admission. The area under the receiver-operating characteristic (AUROC) curve for the prediction of bacteremia via procalcitonin was 0.782 (95% CI 0.779-0.787). The prediction modeling of clinical factors with or without PCT resulted in a similar performance to PCT alone. However, the clinically predicted risk of bacteremia stratified by PCT thresholds allowed the targeting of high-incidence bacteremia groups (e.g., ≥50% positivity). The combined use of PCT and clinical characteristics could be useful in diagnostic stewardship by targeting further advanced diagnostic testing in patients with a high predicted probability of bacteremia.

The Network Relative Model Accuracy (NeRMA) Score can quantify the relative accuracy of prediction models in concurrent external validations

BACKGROUND

Network meta-analysis (NMA) quantifies the relative efficacy of three or more interventions from trials evaluating some, but usually not all, treatments. This study applied the analytical approach of NMA to quantify the relative accuracy of prediction models with distinct patient applicability that are evaluated on the same population ('concurrent external validation').

METHODS

We simulated binary events in 5000 patients using a known risk function. We biased the risk function and modified its precision by pre-specified amounts to create 15 prediction models with varying accuracy and distinct patient applicability. Prediction model accuracy was measured using the Scaled Brier Score (SBS). Overall prediction model accuracy was measured using fixed-effects methods accounting for distinct model applicability patterns. Prediction model accuracy was summarized as the Network Relative Model Accuracy (NeRMA) Score which increases as models become more accurate and ranges from <0 (model less accurate than random guessing) through 0 (accuracy of random guessing) to 1 (most accurate model in concurrent external validation).

RESULTS

The unbiased prediction model had the highest SBS. The NeRMA score correctly ranked all simulated prediction models by the extent of bias from the known risk function. A SAS macro and R-function was created and available to implement the NeRMA Score.

CONCLUSIONS

The NeRMA Score makes it possible to quantify the relative accuracy of binomial prediction models with distinct applicability in a concurrent external validation.

Concurrent external validation of bloodstream infection probability models

OBJECTIVE

Accurately estimating the likelihood of bloodstream infection (BSI) can help clinicians make diagnostic and therapeutic decisions. Many multivariate models predicting BSI probability have been published. This study measured the performance of BSI probability models within the same patient sample.

METHODS

We retrieved validated BSI probability models included in a recently published systematic review that returned a patient-level BSI probability for adults. Model applicability, discrimination, and accuracy was measured in a simple random sample of 4485 admitted adults having blood cultures ordered in the emergency department or the initial 48 hours of hospitalization.

RESULTS

Ten models were included (publication years 1991-2015). Common methodological threats to model performance included overfitting and continuous variable categorization. Restrictive inclusion criteria caused seven models to apply to <15% of validation patients. Model discrimination was less than originally reported in derivation groups (median c-statistic 60%, range 48-69). The observed BSI risk frequently deviated from expected (median integrated calibration index 4.0%, range 0.8-12.4). Notable disagreement in expected BSI probabilities was seen between models (median (25th-75th percentile) relative difference between expected risks 68.0% (28.6-113.6%)).

DISCUSSION

In a large randomly selected external validation population, many published BSI probability models had restricted applicability, limited discrimination and calibration, and extensive inter-model disagreement. Direct comparison of model performance is hampered by dissimilarities between model-specific validation groups.

Bacteremia Prediction With Prognostic Scores and a Causal Probabilistic Network - A Cohort Study of Emergency Department Patients

BACKGROUND

Physicians tend to overestimate patients' pretest probability of having bacteremia. The low yield of blood cultures and contaminants is associated with significant financial cost, as well as increased length of stay and unnecessary antibiotic treatment.

OBJECTIVE

This study examined the abilities of the National Early Warning Score (NEWS), the Quick Sequential Organ Failure Assessment (qSOFA), the Modified Sequential Organ Failure Assessment (mSOFA), and two versions of the causal probabilistic network, SepsisFinder™ (SF) to predict bacteremia in adult emergency department (ED) patients.

METHODS

This cohort study included adult ED patients from a large urban, academic tertiary hospital, with blood cultures obtained within 24 h of admission between 2016 and 2017. The outcome measure was true bacteremia. NEWS, qSOFA, mSOFA, and the two versions of SF score were calculated for all patients based on the first available full set of vital signs within 2 h and laboratory values within 6 h after drawing the blood cultures. Area under the receiver operating characteristic curve (AUROC) was calculated for each scoring system.

RESULTS

The study included 3106 ED patients, of which 199 (6.4%) patients had true bacteremia. The AUROCs for prediction of bacteremia were: NEWS = 0.65, qSOFA = 0.60, SF I = 0.65, mSOFA = 0.71, and SF II = 0.80.

CONCLUSIONS

Scoring systems using only vital signs, NEWS, and SF I showed moderate abilities in predicting bacteremia, whereas qSOFA performed poorly. Scoring systems using both vital signs and laboratory values, mSOFA and especially SF II, showed good abilities in predicting bacteremia.

Usefulness of the 5MPB-Toledo model to predict bacteremia in patients with urinary tract infections in the emergency department

OBJECTIVE

To analyze the usefulness of a new predictive model of bacteremia (5MPB-Toledo) in patients treated for urinary tract infection (UTI) in the emergency department (ED).

METHODS

Prospective and multicenter observational cohort study of the blood cultures (BC) ordered for patients with UTIs in 65 Spanish ED from November 1, 2019, to March 31, 2020. The predictive ability of the model was analyzed with the area under the Receiver Operating Characteristic curve (AUC-ROC). The diagnostic performance was calculated with the chosen cut-off point for sensitivity, specificity, positive predictive value, and negative predictive value.

RESULTS

A total of 1,499 blood cultures were evaluated. True cases of bacteremia were confirmed in 277 (18.5%). The remaining 1,222 cultures (81.5%) were negative. Ninety-four (6.3%) were considered contaminated. The model's area under the ROC curve was 0.937 (95% CI, 0.926-0.949). The prognostic performance with a model's cut-off value of ≥5 points achieved 97.47% (95% CI, 94.64-98.89) sensitivity, 76.68% (95% CI, 74.18-79.00) specificity, 48.65% (95% CI, 44.42-52.89) positive predictive value and 99.26% (95% CI, 98.41-99.67) negative predictive value.

CONCLUSION

The 5MPB-Toledo score is useful for predicting bacteremia in patients with UTIs who visit the ED.

ESTIMATING THE RISK OF BACTERAEMIA IN HOSPITALISED PATIENTS WITH PNEUMOCOCCAL PNEUMONIA

Objective To construct a prediction model for bacteraemia in patients with pneumococcal community-acquired pneumonia (P-CAP) based on variables easily obtained at hospital admission. MethodsThis prospective observational multicentre derivation-validation study was conducted in patients hospitalised with P-CAP between 2000-2020. All cases were diagnosed based on positive urinary antigen tests in the emergency department and had blood cultures taken on admission. A risk score to predict bacteraemia was developed. Results We included 1783 patients with P-CAP (1195 in the derivation and 588 in the validation cohort). A third (33.3%) of the patients had bacteraemia. In the multivariate analysis, the following were identified as independent factors associated with bacteraemia: no influenza vaccination the last year, no pneumococcal vaccination in the last 5 years, blood urea nitrogen (BUN) ≥30 mg/dL, sodium <130 mmol/L, lymphocyte count <800/µl, C-reactive protein ≥200 mg/L, respiratory failure, pleural effusion and no antibiotic treatment before admission. The score yielded good discrimination (AUC 0.732; 95% CI: 0.695-0.769) and calibration (Hosmer-Lemeshow p-value 0.801), with similar performance in the validation cohort (AUC 0.764; 95% CI:0.719-0.809). Conclusions We found nine predictive factors easily obtained on hospital admission that could help achieve early identification of bacteraemia. The prediction model provides a useful tool to guide diagnostic decisions.

Diagnostic stewardship for blood cultures in the emergency department: A multicenter validation and prospective evaluation of a machine learning prediction tool

Background

Overuse of blood cultures (BCs) in emergency departments (EDs) leads to low yields and high numbers of contaminated cultures, accompanied by increased diagnostics, antibiotic usage, prolonged hospitalization, and mortality. We aimed to simplify and validate a recently developed machine learning model to help safely withhold BC testing in low-risk patients.

Methods

We extracted data from the electronic health records (EHR) for 44.123 unique ED visits with BC sampling in the Amsterdam UMC (locations VUMC and AMC; the Netherlands), Zaans Medical Center (ZMC; the Netherlands), and Beth Israel Deaconess Medical Center (BIDMC; United States) in periods between 2011 and 2021. We trained a machine learning model on the VUMC data to predict blood culture outcomes and validated it in the AMC, ZMC, and BIDMC with subsequent real-time prospective evaluation in the VUMC.

Findings

The model had an Area Under the Receiver Operating Characteristics curve (AUROC) of 0.81 (95%-CI = 0.78–0.83) in the VUMC test set. The most important predictors were temperature, creatinine, and C-reactive protein. The AUROCs in the validation cohorts were 0.80 (AMC; 0.78–0.82), 0.76 (ZMC; 0.74–0.78), and 0.75 (BIDMC; 0.74–0.76). During real-time prospective evaluation in the EHR of the VUMC, it reached an AUROC of 0.76 (0.71–0.81) among 590 patients with BC draws in the ED. The prospective evaluation showed that the model can be used to safely withhold blood culture analyses in at least 30% of patients in the ED.

Interpretation

We developed a machine learning model to predict blood culture outcomes in the ED, which retained its performance during external validation and real-time prospective evaluation. Our model can identify patients at low risk of having a positive blood culture. Using the model in practice can significantly reduce the number of blood culture analyses and thus avoid the hidden costs of false-positive culture results.

Funding

This research project was funded by the Amsterdam Public Health – Quality of Care program and the Dutch “Doen of Laten” project (project number: 839205002).

Validation of a predictive model for bacteraemia (MPB5-Toledo) in the patients seen in emergency departments due to infections

OBJECTIVE

To validate a simple risk score to predict bacteremia (MPB5-Toledo) in patients seen in the emergency departments (ED) due to infections.

METHODS

Prospective and multicenter observational cohort study of the blood cultures (BC) ordered in 74 Spanish ED for adults (aged 18 or older) seen from October 1, 2019, to February 29, 2020. The predictive ability of the model was analyzed with the area under the Receiver Operating Characteristic curve (AUC-ROC). The prognostic performance for true bacteremia was calculated with the cut-off values chosen for getting the sensitivity, specificity, positive predictive value and negative predictive value.

RESULTS

A total of 3.843 blood samples wered cultured. True cases of bacteremia were confirmed in 839 (21.83%). The remaining 3.004 cultures (78.17%) were negative. Among the negative, 172 (4.47%) were judged to be contaminated. Low risk for bacteremia was indicated by a score of 0-2 points, intermediate risk by 3-5 points, and high risk by 6-8 points. Bacteremia in these 3 risk groups was predicted for 1.5%, 16.8%, and 81.6%, respectively. The model's area under the receiver operating characteristic curve was 0.930 (95% CI, 0.916-0.948). The prognostic performance with a model's cut-off value of ≥5 points achieved 94.76% (95% CI: 92.97-96.12) sensitivity, 81.56% (95% CI: 80.11-82.92) specificity, and negative predictive value of 98.24% (95% CI: 97.62-98.70).

CONCLUSION

The 5MPB-Toledo score is useful for predicting bacteremia in patients attended in hospital emergency departments for infection.

Using machine learning to predict blood culture outcomes in the emergency department: a single-centre, retrospective, observational study

OBJECTIVES

To develop predictive models for blood culture (BC) outcomes in an emergency department (ED) setting.

DESIGN

Retrospective observational study.

SETTING

ED of a large teaching hospital in the Netherlands between 1 September 2018 and 24 June 2020.

PARTICIPANTS

Adult patients from whom BCs were collected in the ED. Data of demographic information, vital signs, administered medications in the ED and laboratory and radiology results were extracted from the electronic health record, if available at the end of the ED visits.

MAIN OUTCOME MEASURES

The primary outcome was the performance of two models (logistic regression and gradient boosted trees) to predict bacteraemia in ED patients, defined as at least one true positive BC collected at the ED.

RESULTS

In 4885 out of 51 399 ED visits (9.5%), BCs were collected. In 598/4885 (12.2%) visits, at least one of the BCs was true positive. Both a gradient boosted tree model and a logistic regression model showed good performance in predicting BC results with area under curve of the receiver operating characteristics of 0.77 (95% CI 0.73 to 0.82) and 0.78 (95% CI 0.73 to 0.82) in the test sets, respectively. In the gradient boosted tree model, the optimal threshold would predict 69% of BCs in the test set to be negative, with a negative predictive value of over 94%.

CONCLUSIONS

Both models can accurately identify patients with low risk of bacteraemia at the ED in this single-centre setting and may be useful to reduce unnecessary BCs and associated healthcare costs. Further studies are necessary for validation and to investigate the potential clinical benefits and possible risks after implementation.

Prediction of bacteremia at the emergency department during triage and disposition stages using machine learning models

INTRODUCTION

Bacteremia is a common but critical condition with high mortality that requires timely and optimal treatment in the emergency department (ED). The prediction of bacteremia at the ED during triage and disposition stages could support the clinical decisions of ED physicians regarding the appropriate treatment course and safe ED disposition. This study developed and validated machine learning models to predict bacteremia in the emergency department during triage and disposition stages.

METHODS

This study enrolled adult patients who visited a single tertiary hospital from 2016 to 2018 and had at least two sets of blood cultures during their ED stay. Demographic information, chief complaint, triage level, vital signs, and laboratory data were used as model predictors. We developed and validated prediction models using 10 variables at the time of ED triage and 42 variables at the time of disposition. The extreme gradient boosting (XGB) model was compared with the random forest and multivariable logistic regression models. We compared model performance by assessing the area under the receiver operating characteristic curve (AUC), test characteristics, and decision curve analysis.

RESULTS

A total of 24,768 patients were included: 16,197 cases were assigned to development, and 8571 cases were assigned to validation. The proportion of bacteremia was 10.9% and 10.4% in the development and validation datasets, respectively. The Triage XGB model (AUC, 0.718; 95% confidence interval (CI), 0.701-0.735) showed acceptable discrimination performance with a sensitivity over 97%. The Disposition XGB model (AUC, 0.853; 95% CI, 0.840-0.866) showed excellent performance and provided the greatest net benefit throughout the range of thresholds probabilities.

CONCLUSIONS

The Triage XGB model could be used to identify patients with a low risk of bacteremia immediately after initial ED triage. The Disposition XGB model showed excellent discriminative performance.

Machine learning identification of specific changes in myeloid cell phenotype during bloodstream infections

The early identification of bacteremia is critical for ensuring appropriate treatment of nosocomial infections in intensive care unit (ICU) patients. The aim of this study was to use flow cytometric data of myeloid cells as a biomarker of bloodstream infection (BSI). An eight-color antibody panel was used to identify seven monocyte and two dendritic cell subsets. In the learning cohort, immunophenotyping was applied to (1) control subjects, (2) postoperative heart surgery patients, as a model of noninfectious inflammatory responses, and (3) blood culture-positive patients. Of the complex changes in the myeloid cell phenotype, a decrease in myeloid and plasmacytoid dendritic cell numbers, increase in CD14⁺CD16⁺ inflammatory monocyte numbers, and upregulation of neutrophils CD64 and CD123 expression were prominent in BSI patients. An extreme gradient boosting (XGBoost) algorithm called the “infection detection and ranging score” (iDAR), ranging from 0 to 100, was developed to identify infection-specific changes in 101 phenotypic variables related to neutrophils, monocytes and dendritic cells. The tenfold cross-validation achieved an area under the receiver operating characteristic (AUROC) of 0.988 (95% CI 0.985–1) for the detection of bacteremic patients. In an out-of-sample, in-house validation, iDAR achieved an AUROC of 0.85 (95% CI 0.71–0.98) in differentiating localized from bloodstream infection and 0.95 (95% CI 0.89–1) in discriminating infected from noninfected ICU patients. In conclusion, a machine learning approach was used to translate the changes in myeloid cell phenotype in response to infection into a score that could identify bacteremia with high specificity in ICU patients.

A bacteraemia risk prediction model: development and validation in an emergency medicine population

OBJECTIVE

Design a risk model to predict bacteraemia in patients attended in emergency departments (ED) for an episode of infection.

METHODS

This was a national, prospective, multicentre, observational cohort study of blood cultures (BC) collected from adult patients (≥ 18 years) attended in 71 Spanish EDs from October 1 2019 to March 31, 2020. Variables with a p value < 0.05 were introduced in the univariate analysis together with those of clinical significance. The final selection of variables for the scoring scale was made by logistic regression with selection by introduction. The results obtained were internally validated by dividing the sample in a derivation and a validation cohort.

RESULTS

A total of 4,439 infectious episodes were included. Of these, 899 (20.25%) were considered as true bacteraemia. A predictive model for bacteraemia was defined with seven variables according to the Bacteraemia Prediction Model of the INFURG-SEMES group (MPB-INFURG-SEMES). The model achieved an area under the curve-receiver operating curve of 0.924 (CI 95%:0.914-0.934) in the derivation cohort, and 0.926 (CI 95%: 0.910-0.942) in the validation cohort. Patients were then split into ten risk categories, and had the following rates of risk: 0.2%(0 points), 0.4%(1 point), 0.9%(2 points), 1.8%(3 points), 4.7%(4 points), 19.1% (5 points), 39.1% (6 points), 56.8% (7 points), 71.1% (8 points), 82.7% (9 points) and 90.1% (10 points). Findings were similar in the validation cohort. The cut-off point of five points provided the best precision with a sensitivity of 95.94%, specificity of 76.28%, positive predictive value of 53.63% and negative predictive value of 98.50%.

CONCLUSION

The MPB-INFURG-SEMES model may be useful for the stratification of risk of bacteraemia in adult patients with infection in EDs, together with clinical judgement and other variables independent of the process and the patient.

A comparison of predictors for mortality and bacteraemia in patients suspected of infection

BACKGROUND

Stratification by clinical scores of patients suspected of infection can be used to support decisions on treatment and diagnostic workup. Seven clinical scores, SepsisFinder (SF), National Early Warning Score (NEWS), Sequential Orgen Failure Assessment (SOFA), Mortality in Emergency Department Sepsis (MEDS), quick SOFA (qSOFA), Shapiro Decision Rule (SDR) and Systemic Inflammatory Response Syndrome (SIRS), were evaluated for their ability to predict 30-day mortality and bacteraemia and for their ability to identify a low risk group, where blood culture may not be cost-effective and a high risk group where direct-from-blood PCR (dfbPCR) may be cost effective.

METHODS

Retrospective data from two Danish and an Israeli hospital with a total of 1816 patients were used to calculate the seven scores.

RESULTS

SF had higher Area Under the Receiver Operating curve than the clinical scores for prediction of mortality and bacteraemia, significantly so for MEDS, qSOFA and SIRS. For mortality predictions SF also had significantly higher area under the curve than SDR. In a low risk group identified by SF, consisting of 33% of the patients only 1.7% had bacteraemia and mortality was 4.2%, giving a cost of € 1976 for one positive result by blood culture. This was higher than the cost of € 502 of one positive dfbPCR from a high risk group consisting of 10% of the patients, where 25.3% had bacteraemia and mortality was 24.2%.

CONCLUSION

This may motivate a health economic study of whether resources spent on low risk blood cultures might be better spent on high risk dfbPCR.

Bacteraemia, antimicrobial susceptibility and treatment among Campylobacter-associated hospitalisations in the Australian Capital Territory: a review

BACKGROUND

Campylobacter spp. cause mostly self-limiting enterocolitis, although a significant proportion of cases require hospitalisation highlighting potential for severe disease. Among people admitted, blood culture specimens are frequently collected and antibiotic treatment is initiated. We sought to understand clinical and host factors associated with bacteraemia, antibiotic treatment and isolate non-susceptibility among Campylobacter-associated hospitalisations.

METHODS

Using linked hospital microbiology and administrative data we identified and reviewed Campylobacter-associated hospitalisations between 2004 and 2013. We calculated population-level incidence for Campylobacter bacteraemia and used logistic regression to examine factors associated with bacteraemia, antibiotic treatment and isolate non-susceptibility among Campylobacter-associated hospitalisations.

RESULTS

Among 685 Campylobacter-associated hospitalisations, we identified 25 admissions for bacteraemia, an estimated incidence of 0.71 cases per 100,000 population per year. Around half of hospitalisations (333/685) had blood culturing performed. Factors associated with bacteraemia included underlying liver disease (aOR 48.89, 95% CI 7.03-340.22, p < 0.001), Haematology unit admission (aOR 14.67, 95% CI 2.99-72.07, p = 0.001) and age 70-79 years (aOR 4.93, 95% CI 1.57-15.49). Approximately one-third (219/685) of admissions received antibiotics with treatment rates increasing significantly over time (p < 0.05). Factors associated with antibiotic treatment included Gastroenterology unit admission (aOR 3.75, 95% CI 1.95-7.20, p < 0.001), having blood cultures taken (aOR 2.76, 95% CI 1.79-4.26, p < 0.001) and age 40-49 years (aOR 2.34, 95% CI 1.14-4.79, p = 0.02). Non-susceptibility of isolates to standard antimicrobials increased significantly over time (p = 0.01) and was associated with overseas travel (aOR 11.80 95% CI 3.18-43.83, p < 0.001) and negatively associated with tachycardia (aOR 0.48, 95%CI 0.26-0.88, p = 0.02), suggesting a healthy traveller effect.

CONCLUSIONS

Campylobacter infections result in considerable hospital burden. Among those admitted to hospital, an interplay of factors involving clinical presentation, presence of underlying comorbidities, complications and increasing age influence how a case is investigated and managed.

An Artificial Intelligence Approach to Bloodstream Infections Prediction

This study aimed to develop an early prediction model for identifying patients with bloodstream infections. The data resource was taken from 2015 to 2019 at Taichung Veterans General Hospital, and a total of 1647 bloodstream infection episodes and 3552 non-bloodstream infection episodes in the intensive care unit (ICU) were included in the model development and evaluation. During the data analysis, 30 clinical variables were selected, including patients’ basic characteristics, vital signs, laboratory data, and clinical information. Five machine learning algorithms were applied to examine the prediction model performance. The findings indicated that the area under the receiver operating characteristic curve (AUROC) of the prediction performance of the XGBoost model was 0.825 for the validation dataset and 0.821 for the testing dataset. The random forest model also presented higher values for the AUROC on the validation dataset and testing dataset, which were 0.855 and 0.851, respectively. The tree-based ensemble learning model enabled high detection ability for patients with bloodstream infections in the ICU. Additionally, the analysis of importance of features revealed that alkaline phosphatase (ALKP) and the period of the central venous catheter are the most important predictors for bloodstream infections. We further explored the relationship between features and the risk of bloodstream infection by using the Shapley Additive exPlanations (SHAP) visualized method. The results showed that a higher prothrombin time is more prominent in a bloodstream infection. Additionally, the impact of a lower platelet count and albumin was more prominent in a bloodstream infection. Our results provide additional clinical information for cut-off laboratory values to assist clinical decision-making in bloodstream infection diagnostics.

The Diagnostic Yield of Routine Admission Blood Cultures in Critically Ill Patients

OBJECTIVES

Although the Surviving Sepsis Campaign bundle recommends obtaining blood cultures within 1 hour of sepsis recognition, adherence is suboptimal in many settings. We, therefore, implemented routine blood culture collection for all nonelective ICU admissions (regardless of infection suspicion) and evaluated its diagnostic yield.

DESIGN

A before-after analysis.

SETTING

A mixed-ICU of a tertiary care hospital in the Netherlands.

PATIENTS

Patients acutely admitted to the ICU between January 2015 and December 2018.

MEASUREMENTS AND MAIN RESULTS

Automatic orders for collecting a single set of blood cultures immediately upon ICU admission were implemented on January 1, 2017. Blood culture results and the impact of contaminated blood cultures were compared for 2015-2016 (before period) and 2017-2018 (after period). Positive blood cultures were categorized as bloodstream infection or contamination. Blood cultures were obtained in 573 of 1,775 patients (32.3%) and in 1,582 of 1,871 patients (84.5%) in the before and after periods, respectively (p < 0.0001), and bloodstream infection was diagnosed in 95 patients (5.4%) and 154 patients (8.2%) in both study periods (relative risk 1.5; 95% CI 1.2-2.0; p = 0.0006). The estimated number needed to culture for one additional patient with bloodstream infection was 17. Blood culture contamination occurred in 40 patients (2.3%) and 180 patients (9.6%) in the before period and after period, respectively (relative risk 4.3; 95% CI 3.0-6.0; p < 0.0001). Rate of vancomycin use or presumed episodes of catheter-related bloodstream infections treated with antibiotics did not differ between both study periods.

CONCLUSIONS

Implementation of routine blood cultures was associated with a 1.5-fold increase of detected bloodstream infection. The 4.3-fold increase in contaminated blood cultures was not associated with an increase in vancomycin use in the ICU.

Diagnostic performance of food consumption for bacteraemia in patients admitted with suspected infection: a prospective cohort study

OBJECTIVES

A previous study reported that food consumption is useful to rule out bacteraemia in hospitalised patients. We aimed to validate the diagnostic performance of (1) food consumption and (2) a previously reported algorithm using food consumption and shaking chills for bacteraemia in patients admitted to hospital with suspected infection.

DESIGN

Prospective cohort study.

SETTING

Department of General Medicine in two acute care hospitals in Japan.

PARTICIPANTS

A total of 2009 adult patients who underwent at least two blood cultures on admission.

PRIMARY OUTCOME MEASURES

The reference standard for bacteraemia was judgement by two independent specialists of infectious diseases. Food consumption was evaluated by the physician in charge asking the patient or their caregivers the following question on admission: 'What percentage of usual food intake were you able to eat during the past 24 hours?'

RESULTS

Among 2009 patients, 326 patients were diagnosed with bacteraemia (16.2%). Diagnostic performance of food consumption was sensitivity of 84.4% (95% CI 80.1 to 88), specificity of 19.8% (95% CI 18 to 21.8), positive predictive value (PPV) of 16.9% (95% CI 15.2 to 18.9) and negative predictive value (NPV) of 86.8% (95% CI 83.1 to 89.8). The discriminative performance was an area under the curve of 0.53 (95% CI 0.50 to 0.56). The performance of the algorithm using food consumption and shaking chills was sensitivity of 89% (95% CI 85.1 to 91.9), specificity of 18.8% (95% CI 17 to 20.7), PPV of 17.5% (95% CI 15.7 to 19.4) and NPV of 89.8% (95% CI 86.2 to 92.5).

CONCLUSION

Our results did not show the usefulness of food consumption and the algorithm using food consumption and shaking chills for the diagnosis of bacteraemia in patients admitted to hospital with suspected infection.

[Usefulness of the 5MPB-Toledo model to predict bacteremia in patients with community-acquired pneumonia in the Emergency Department]

OBJECTIVE

To analyse a new risk score to predict bacteremia in the patients with Community-acquired Pneumonia (CAP) in the emergency departments.

METHODS

Prospective and multicenter observational cohort study of the blood cultures ordered in 74 Spanish emergency departments for patients with CAP seen from November 1, 2019, to March 31, 2020. The predictive ability of the model was analyzed with the area under the Receiver Operating Characteristic curve (AUC-ROC). The prognostic performance for true bacteremia was calculated with the chosen cut-off for getting the sensitivity, specificity, positive predictive value and negative predictive value.

RESULTS

A total of 1,020 blood samples wered cultured. True cases of bacteremia were confirmed in 162 (15.9%). The remaining 858 cultures (84.1%) wered negative. And, 59 (5.8%) were judged to be contaminated. The model´s area under the receiver operating characteristic curve was 0.915 (95% CI, 0.898-0.933). The prognostic performance with a model´s cut-off value of ≥ 5 points achieved 97.5% (95% CI, 95.1-99.9) sensitivity, 73.2% (95% CI, 70.2-76.2) specificity, 40.9% (95% CI, 36.4-45.1) positive predictive value and 99.4% (95% CI, 99.1-99.8) negative predictive value.

CONCLUSIONS

The 5MPB-Toledo score is useful for predicting bacteremia in the patients with CAP seen in the emergency departments.

The Development and Validation of a Machine Learning Model to Predict Bacteremia and Fungemia in Hospitalized Patients Using Electronic Health Record Data

OBJECTIVES

Bacteremia and fungemia can cause life-threatening illness with high mortality rates, which increase with delays in antimicrobial therapy. The objective of this study is to develop machine learning models to predict blood culture results at the time of the blood culture order using routine data in the electronic health record.

DESIGN

Retrospective analysis of a large, multicenter inpatient data.

SETTING

Two academic tertiary medical centers between the years 2007 and 2018.

SUBJECTS

All hospitalized patients who received a blood culture during hospitalization.

INTERVENTIONS

The dataset was partitioned temporally into development and validation cohorts: the logistic regression and gradient boosting machine models were trained on the earliest 80% of hospital admissions and validated on the most recent 20%.

MEASUREMENTS AND MAIN RESULTS

There were 252,569 blood culture days-defined as nonoverlapping 24-hour periods in which one or more blood cultures were ordered. In the validation cohort, there were 50,514 blood culture days, with 3,762 cases of bacteremia (7.5%) and 370 cases of fungemia (0.7%). The gradient boosting machine model for bacteremia had significantly higher area under the receiver operating characteristic curve (0.78 [95% CI 0.77-0.78]) than the logistic regression model (0.73 [0.72-0.74]) (p < 0.001). The model identified a high-risk group with over 30 times the occurrence rate of bacteremia in the low-risk group (27.4% vs 0.9%; p < 0.001). Using the low-risk cut-off, the model identifies bacteremia with 98.7% sensitivity. The gradient boosting machine model for fungemia had high discrimination (area under the receiver operating characteristic curve 0.88 [95% CI 0.86-0.90]). The high-risk fungemia group had 252 fungemic cultures compared with one fungemic culture in the low-risk group (5.0% vs 0.02%; p < 0.001). Further, the high-risk group had a mortality rate 60 times higher than the low-risk group (28.2% vs 0.4%; p < 0.001).

CONCLUSIONS

Our novel models identified patients at low and high-risk for bacteremia and fungemia using routinely collected electronic health record data. Further research is needed to evaluate the cost-effectiveness and impact of model implementation in clinical practice.

Characteristics of spirochetemic patients with a solitary erythema migrans skin lesion in Europe

Neither pre-treatment characteristics, nor the outcome after antibiotic therapy, have been reported for spirochetemic European patients with Lyme borreliosis. In the present study, patients with a solitary erythema migrans (EM) who had a positive blood culture for either Borrelia afzelii (n = 116) or Borrelia garinii (n = 37) were compared with age- and sex-matched patients who had a negative blood culture, but were culture positive for the corresponding Borrelia species from skin. Collectively, spirochetemic patients significantly more often recalled a tick bite at the site of the EM skin lesion, had a shorter time interval from the bite to the onset of EM, had a shorter duration of the skin lesion prior to diagnosis, and had a smaller EM skin lesion that was more often homogeneous in appearance. Similar results were found for the subset of spirochetemic patients infected with B. afzelii but not for those infected with B. garinii. However, patients with B. garinii bacteremia had faster-spreading and larger EM skin lesions, and more often reported itching at the site of the lesion than patients with B. afzelii bacteremia. Treatment failures were rare (7/306 patients, 2.3%) and were not associated with having spirochetemia or with which Borrelia species was causing the infection.

Added value of clinical prediction rules for bacteremia in hemodialysis patients: An external validation study

Introduction

Having developed a clinical prediction rule (CPR) for bacteremia among hemodialysis (HD) outpatients (BAC-HD score), we performed external validation.

Materials & methods

Data were collected on maintenance HD patients at two Japanese tertiary-care hospitals from January 2013 to December 2015. We enrolled 429 consecutive patients (aged ≥ 18 y) on maintenance HD who had had two sets of blood cultures drawn on admission to assess for bacteremia. We validated the predictive ability of the CPR using two validation cohorts. Index tests were the BAC-HD score and a CPR developed by Shapiro et al. The outcome was bacteremia, based on the results of the admission blood cultures. For added value, we also measured changes in the area under the receiver operating characteristic curve (AUC) using logistic regression and Net Reclassification Improvement (NRI), in which each CPR was added to the basic model.

Results

In Validation cohort 1 (360 subjects), compared to a Model 1 (Basic Model) AUC of 0.69 (95% confidence interval [95% CI]: 0.59–0.80), the AUC of Model 2 (Basic model + BAC-HD score) and Model 3 (Basic model + Shapiro’s score) increased to 0.8 (95% CI: 0.71–0.88) and 0.73 (95% CI: 0.63–0.83), respectively. In validation cohort 2 (96 subjects), compared to a Model 1 AUC of 0.81 (95% CI: 0.68–0.94), the AUCs of Model 2 and Model 3 increased to 0.83 (95% CI: 0.72–0.95) and 0.85 (95% CI: 0.76–0.94), respectively. NRIs on addition of the BAC-HD score and Shapiro’s score were 0.3 and 0.06 in Validation cohort 1, and 0.27 and 0.13, respectively, in Validation cohort 2.

Conclusion

Either the BAC-HD score or Shapiro’s score may improve the ability to diagnose bacteremia in HD patients. Reclassification was better with the BAC-HD score.

Validation of a predictive model for bacteraemia (MPB5-Toledo) in the patients seen in emergency departments due to infections

OBJECTIVE

To validate a simple risk score to predict bacteremia (MPB5-Toledo) in patients seen in the emergency departments (ED) due to infections.

METHODS

Prospective and multicenter observational cohort study of the blood cultures (BC) ordered in 74 Spanish ED for adults (aged 18 or older) seen from from October 1, 2019, to February 29, 2020. The predictive ability of the model was analyzed with the area under the Receiver Operating Characteristic curve (AUC-ROC). The prognostic performance for true bacteremia was calculated with the cut-off values chosen for getting the sensitivity, specificity, positive predictive value and negative predictive value.

RESULTS

A total of 3.843 blood samples wered cultured. True cases of bacteremia were confirmed in 839 (21.83%). The remaining 3.004 cultures (78.17%) were negative. Among the negative, 172 (4.47%) were judged to be contaminated. Low risk for bacteremia was indicated by a score of 0 to 2 points, intermediate risk by 3 to 5 points, and high risk by 6 to 8 points. Bacteremia in these 3 risk groups was predicted for 1.5%, 16.8%, and 81.6%, respectively. The model's area under the receiver operating characteristic curve was 0.930 (95% CI, 0.916-0.948). The prognostic performance with a model's cut-off value of ≥ 5 points achieved 94.76% (95% CI: 92.97-96.12) sensitivity, 81.56% (95% CI: 80.11-82.92) specificity, and negative predictive value of 98.24% (95% CI: 97.62-98.70).

CONCLUSION

The 5MPB-Toledo score is useful for predicting bacteremia in patients attended in hospital emergency departments for infection.

Risk Factors for Mortality, Intensive Care Unit Admission, and Bacteremia in Patients Suspected of Sepsis at the Emergency Department: A Prospective Cohort Study

Background

There is a clear need for a better assessment of independent risk factors for in-hospital mortality, intensive care unit admission, and bacteremia in patients presenting with suspected sepsis at the emergency department.

Methods

A prospective observational cohort study including 1690 patients was performed. Two multivariable logistic regression models were used to identify independent risk factors.

Results

Sequential organ failure assessment (SOFA) score of ≥2 and serum lactate of ≥2mmol/L were associated with all outcomes. Other independent risk factors were individual SOFA variables and systemic inflammatory response syndrome variables but varied per outcome. Mean arterial pressure <70 mmHg negatively impacted all outcomes.

Conclusions

These readily available measurements can help with early risk stratification and prediction of prognosis.

Blood urea nitrogen to serum albumin ratio independently predicts 30-day mortality and severity in patients with Escherichia coli bacteraemia

BACKGROUND

Elevated blood urea nitrogen to serum albumin (BUN/ALB) ratio had been identified as an independent risk factor related to mortality in community-acquired and hospital-acquired pneumonia. This study aimed to investigate whether this clinical index can predict the clinical outcomes of E. coli bacteraemia.

MATERIAL AND METHODS

Clinical data were collected from patients with E. coli bacteraemia attended at our hospital between January 2012 and December 2018. The endpoints were mortality within 30 days after the diagnosis of E. coli bacteraemia and intensive care (IC) requirement. Cox regression analysis was performed to evaluate the risk factors.

RESULTS

A total of 398 patients with E. coli bacteraemia were enrolled in this study and 56 patients died within 30 days after bacteraemia onset. Multivariate Cox regression analysis showed that age greater than 65 years, lymphocyte count<.8×10e9/L, elevated BUN/ALB ratio, increased SOFA score, carbapenem resistance, central venous catheterization before onset of bacteraemia, and infection originating from abdominal cavity were independent risk factors for 30-day mortality (P<.05). The risk factors associated with IC requirement were similar to those for 30-day mortality except central venous catheterization before onset of bacteraemia. The area under the receiver-operating characteristic curve for BUN/ALB ratio predicting 30-day mortality and IC requirement was similar to that for SOFA score, but higher than that for lymphocyte count. The cut-off points of BUN/ALB ratio to predict 30-day mortality and IC requirement were both .3.

CONCLUSIONS

BUN/ALB ratio is a simple but independent predictor of 30-day mortality and severity in E. coli bacteraemia. A higher BUN/ALB ratio at the onset of bacteraemia predicts a higher mortality rate and IC requirement.

Risk of bacteremia in patients presenting with shaking chills and vomiting - a prospective cohort study

Chills and vomiting have traditionally been associated with severe bacterial infections and bacteremia. However, few modern studies have in a prospective way evaluated the association of these signs with bacteremia, which is the aim of this prospective, multicenter study. Patients presenting to the emergency department with at least one affected vital sign (increased respiratory rate, increased heart rate, altered mental status, decreased blood pressure or decreased oxygen saturation) were included. A total of 479 patients were prospectively enrolled. Blood cultures were obtained from 197 patients. Of the 32 patients with a positive blood culture 11 patients (34%) had experienced shaking chills compared with 23 (14%) of the 165 patients with a negative blood culture, P = 0.009. A logistic regression was fitted to show the estimated odds ratio (OR) for a positive blood culture according to shaking chills. In a univariate model shaking chills had an OR of 3.23 (95% CI 1.35–7.52) and in a multivariate model the OR was 5.9 (95% CI 2.05–17.17) for those without prior antibiotics adjusted for age, sex, and prior antibiotics. The presence of vomiting was also addressed, but neither a univariate nor a multivariate logistic regression showed any association between vomiting and bacteremia. In conclusion, among patients at the emergency department with at least one affected vital sign, shaking chills but not vomiting were associated with bacteremia.

Early diagnosis of bloodstream infections in the intensive care unit using machine-learning algorithms

PURPOSE

We aimed to develop a machine-learning (ML) algorithm that can predict intensive care unit (ICU)-acquired bloodstream infections (BSI) among patients suspected of infection in the ICU.

METHODS

The study was based on patients' electronic health records at Beth Israel Deaconess Medical Center (BIDMC) in Boston, Massachusetts, USA, and at Rambam Health Care Campus (RHCC), Haifa, Israel. We included adults from whom blood cultures were collected for suspected BSI at least 48 h after admission. Clinical data, including time-series variables and their interactions, were analyzed by an ML algorithm at each site. Prediction ability for ICU-acquired BSI was assessed by the area under the receiver operating characteristics (AUROC) of ten-fold cross-validation and validation sets with 95% confidence intervals.

RESULTS

The datasets comprised 2351 patients from BIDMC (151 with BSI) and 1021 from RHCC (162 with BSI). The median (inter-quartile range) age was 62 (51-75) and 56 (38-69) years, respectively; the median Acute Physiology and Chronic Health Evaluation II scores were 26 (21-32) and 24 (20-29), respectively. The means of the cross-validation AUROCs were 0.87 ± 0.02 for BIDMC and 0.93 ± 0.03 for RHCC. AUROCs of 0.89 ± 0.01 and 0.92 ± 0.02 were maintained in both centers with internal validation, while external validation deteriorated. Valuable predictors were mainly the trends of time-series variables such as laboratory results and vital signs.

CONCLUSION

An ML approach that uses temporal and site-specific data achieved high performance in recognizing BC samples with a high probability for ICU-acquired BSI.

Practical Guidance for Clinical Microbiology Laboratories: A Comprehensive Update on the Problem of Blood Culture Contamination and a Discussion of Methods for Addressing the Problem

In this review, we present a comprehensive discussion of matters related to the problem of blood culture contamination. Issues addressed include the scope and magnitude of the problem, the bacteria most often recognized as contaminants, the impact of blood culture contamination on clinical microbiology laboratory function, the economic and clinical ramifications of contamination, and, perhaps most importantly, a systematic discussion of solutions to the problem. We conclude by providing a series of unanswered questions that pertain to this important issue.

[Predictive factors of bacteraemia in the patients seen in emergency departments due to infections]

Objetivos

Analizar los factores predictivos de bacteriemia en los pacientes atendidos en el servicio de urgencias (SU) por un episodio de infección.

Pacientes y métodos

Estudio observacional, retrospectivo, descriptivo y analítico de todos los hemocultivos extraídos en un SU en los pacientes adultos (≥ 18 años) atendidos por infección desde el 1-1-2018 hasta el 1-7-2018. Se realizó seguimiento durante 30 días. Se analizaron 38 variables independientes (epidemiológicas, de comorbilidad, funcionales, clínicas y analíticas) que pudieran predecir la existencia de bacteriemia. Se realizó un estudio univariado y multivariante mediante regresión logística.

Resultados

Se incluyeron 1.425 episodios de hemocultivos extraídos. De ellos se consideraron como bacteriemias verdaderas 179 (12,6 %) y como HC negativos 1.246 (87,4 %). Entre los negativos, 1.130 (79,3%) no tuvieron crecimiento y 116 (8,1%) se consideraron contaminados. Cinco variables se asociaron de forma significativa como predictoras de bacteriemia verdadera: procalcitonina (PCT) sérica ≥ 0,51 ng/ml [odds ratio (OR): 4,52; intervalo de confianza (IC) al 95%: 4,20-4,84; p <0,001], temperatura > 38,3°C [OR: 1,60; IC al 95%: 1,29-1,90; p <0,001], presión arterial sistólica (PAS) < 100 mmHg [OR: 3,68; IC al 95%: 2,78-4,58; p <0,001], shock séptico [OR: 2,96; IC al 95%: 1,78-4,13; p <0,001] y la existencia de neoplasia [OR: 1,73; IC al 95%: 1,27-2,20; p <0,001].

Conclusiones

. Existen varios factores disponibles tras una primera valoración en el SU, entre ellos la PCT sérica, la temperatura, la hipotensión con/sin criterios de shock séptico y la existencia de neoplasia, que predicen la existencia de bacteriemia verdadera.

ESKAPE Pathogens in Bloodstream Infections Are Associated With Higher Cost and Mortality but Can Be Predicted Using Diagnoses Upon Admission

Background

ESKAPE bacteria are thought to be especially resistant to antibiotics, and their resistance and prevalence in bloodstream infections are rising. Large studies are needed to better characterize the clinical impact of these bacteria and to develop algorithms that alert clinicians when patients are at high risk of an ESKAPE infection.

Methods

From a US data set of >1.1 M patient encounters, we evaluated if ESKAPE pathogens produced worse outcomes than non-ESKAPE pathogens and if an ESKAPE infection could be predicted using simple word group algorithms built from decision trees.

Results

We found that ESKAPE pathogens represented 42.2% of species isolated from bloodstream infections and, compared with non-ESKAPE pathogens, were associated with a 3.3-day increase in length of stay, a $5500 increase in cost of care, and a 2.1% absolute increase in mortality (P < 1e-99). ESKAPE pathogens were not universally more resistant to antibiotics, but only to select antibiotics (P < 5e-6), particularly against common empiric therapies. In addition, simple word group algorithms predicted ESKAPE pathogens with a positive predictive value of 7.9% to 56.2%, exceeding 4.8% by random guessing (P < 1e-99).

Conclusions

Taken together, these data highlight the pathogenicity of ESKAPE bacteria, potential mechanisms of their pathogenicity, and the potential to predict ESKAPE infections upon admission. Implementing word group algorithms could enable earlier and targeted therapies against ESKAPE bacteria and thus reduce their burden on the health care system.

Best Practices of Blood Cultures in Low- and Middle-Income Countries

Bloodstream infections (BSI) have a substantial impact on morbidity and mortality worldwide. Despite scarcity of data from many low- and middle-income countries (LMICs), there is increasing awareness of the importance of BSI in these countries. For example, it is estimated that the global mortality of non-typhoidal Salmonella bloodstream infection in children under 5 already exceeds that of malaria. Reliable and accurate diagnosis of these infections is therefore of utmost importance. Blood cultures are the reference method for diagnosis of BSI. LMICs face many challenges when implementing blood cultures, due to financial, logistical, and infrastructure-related constraints. This review aims to provide an overview of the state-of-the-art of sampling and processing of blood cultures, with emphasis on its use in LMICs. Laboratory processing of blood cultures is relatively straightforward and can be done without the need for expensive and complicated equipment. Automates for incubation and growth monitoring have become the standard in high-income countries (HICs), but they are still too expensive and not sufficiently robust for imminent implementation in most LMICs. Therefore, this review focuses on "manual" methods of blood culture, not involving automated equipment. In manual blood cultures, a bottle consisting of a broth medium supporting bacterial growth is incubated in a normal incubator and inspected daily for signs of growth. The collection of blood for blood culture is a crucial step in the process, as the sensitivity of blood cultures depends on the volume sampled; furthermore, contamination of the blood culture (accidental inoculation of environmental and skin bacteria) can be avoided by appropriate antisepsis. In this review, we give recommendations regarding appropriate blood culture sampling and processing in LMICs. We present feasible methods to detect and speed up growth and discuss some challenges in implementing blood cultures in LMICs, such as the biosafety aspects, supply chain and waste management.

Clinical- vs. model-based selection of patients suspected of sepsis for direct-from-blood rapid diagnostics in the emergency department: a retrospective study

Selecting high-risk patients may improve the cost-effectiveness of rapid diagnostics. Our objective was to assess whether model-based selection or clinical selection is better for selecting high-risk patients with a high rate of bacteremia and/or DNAemia. This study involved a model-based, retrospective selection of patients from a cohort from which clinicians selected high-risk patients for rapid direct-from-blood diagnostic testing. Patients were included if they were suspected of sepsis and had blood cultures ordered at the emergency department. Patients were selected by the model by adding those with the highest probability of bacteremia until the number of high-risk patients selected by clinicians was reached. The primary outcome was bacteremia rate. Secondary outcomes were DNAemia rate, and 30-day mortality. Data were collected for 1395 blood cultures. Following exclusion, 1142 patients were included in the analysis. In each high-risk group, 220/1142 were selected, where 55 were selected both by clinicians and the model. For the remaining 165 in each group, the model selected for a higher bacteremia rate (74/165, 44.8% vs. 45/165, 27.3%, p = 0.001), and a higher 30-day mortality (49/165, 29.7% vs. 19/165, 11.5%, p = 0.00004) than the clinically selected group. The model outperformed clinicians in selecting patients with a high rate of bacteremia. Using such a model for risk stratification may contribute towards closing the gap in cost between rapid and culture-based diagnostics.

Machine learning for fast identification of bacteraemia in SIRS patients treated on standard care wards: a cohort study

Bacteraemia is a life-threating condition requiring immediate diagnostic and therapeutic actions. Blood culture (BC) analyses often result in a low true positive result rate, indicating its improper usage. A predictive model might assist clinicians in deciding for whom to conduct or to avoid BC analysis in patients having a relevant bacteraemia risk. Predictive models were established by using linear and non-linear machine learning methods. To obtain proper data, a unique data set was collected prior to model estimation in a prospective cohort study, screening 3,370 standard care patients with suspected bacteraemia. Data from 466 patients fulfilling two or more systemic inflammatory response syndrome criteria (bacteraemia rate: 28.8%) were finally used. A 29 parameter panel of clinical data, cytokine expression levels and standard laboratory markers was used for model training. Model tuning was performed in a ten-fold cross validation and tuned models were validated in a test set (80:20 random split). The random forest strategy presented the best result in the test set validation (ROC-AUC: 0.729, 95%CI: 0.679–0.779). However, procalcitonin (PCT), as the best individual variable, yielded a similar ROC-AUC (0.729, 95%CI: 0.679–0.779). Thus, machine learning methods failed to improve the moderate diagnostic accuracy of PCT.

Evaluation of a model to improve collection of blood cultures in patients with sepsis in the emergency room

Sepsis begins outside of the hospital for nearly 80% of patients and the emergency room (ER) represents the first contact with the health care system. This study evaluates a project to improve collection of blood cultures (BCs) in patients with sepsis in the ER consisting of staff education and completion of the appropriate BC pre-analytical phase. A retrospective observational study performed to analyse the data on BC collection in the ER before and after a three-phase project. The first phase (1 January to 30 June 2015) before the intervention consisted of evaluation of data on BCs routinely collected in the ER. The second phase (1 July to 31 December 2015) was the intervention phase in which educational courses on sepsis recognition and on pre-analytical phase procedures (including direct incubation) were provided to ER staff. The third phase (1 January to 30 June 2016; after the intervention) again consisted of evaluation. Before the intervention, out of 24,738 admissions to the ER, 103 patients (0.4%) were identified as septic and had BCs drawn (359 BC bottles); 19 out of 103 patients (18.4%) had positive BCs. After the intervention, out of 24,702 admissions, 313 patients (1.3%) had BCs drawn (1,242 bottles); of these, 96 (30.7%) had positive BCs. Comparing the first and third periods, an increase in the percentage of patients with BCs collected (from 0.4% to 1.3% respectively, p < 0.0001) and an increase in the percentages of patients with true-positive BCs (from 0.08% to 0.39% of all patients evaluated respectively, p < 0.0001) were observed. The isolation of bacteria by BCs increased 3.25-fold after project implementation. These results can be principally ascribed to an improved awareness of sepsis in the staff associated with improved pre-analytical phase procedures in BC collection.

Development and Validation of a Clinical Prediction Rule for Bacteremia among Maintenance Hemodialysis Patients in Outpatient Settings

Background

To our knowledge, no reliable clinical prediction rule (CPR) for identifying bacteremia in hemodialysis (HD) patients has been established. The aim of this study was to develop a CPR for bacteremia in maintenance HD patients visiting the outpatient department.

Methods

This multicenter cohort study involved consecutive maintenance HD patients who visited the outpatient clinic or emergency room of seven Japanese institutions between August 2011 and July 2013. The outcome measure was bacteremia diagnosed based on the results of blood cultures. The candidate predictors for bacteremia were extracted through a literature review. A CPR for bacteremia was developed using a coefficient-based multivariable logistic regression scoring method, and calibration was performed. The test performance was then assessed for the CPR.

Results

Of 507 patients eligible for the study, we analyzed the 293 with a complete dataset for candidate predictors. Of these 293 patients, 48 (16.4%) were diagnosed with bacteremia. At the conclusion of the deviation process, body temperature ≥ 38.3°C, heart rate ≥ 125 /min, C-reactive protein ≥ 10 mg/dL, alkaline phosphatase >360 IU/L, and no prior antibiotics use within the past week were retained and scored. The CPR had a good fit for the model on calibration. The AUC of the CPR was 0.76, and for score CPR ≥ 2, the sensitivity and specificity were 89.6% and 51.4%, respectively.

Conclusions

We established a simple CPR for bacteremia in maintenance HD patients using routinely obtained clinical information in an outpatient setting. This model may facilitate more appropriate clinical decision making.